Morelli, co-founder of bitdrift and a former Lyft engineering leader, was asked what people get wrong about observability costs. He said companies spend huge sums on telemetry that nobody reads, which is why his company prices per active device rather than per data ingested.
…about user about per device does sort of like share the success of a tool that's spreading through a company. But give us a sense of what people do actually get wrong about observability costs. Like what do they not know when they start building up a observability team and investing in these tools?
Yeah. Well, I think it's like absolutely critical.
Like I think even now with more software, more complexity being built, sometimes stuff that humans aren't looking at, you need observability even more to understand what actually is happening. So we call it actual reality, right? It's very easy to test the net and simulate and do evals internally. But when it hits the real world, right, like that's that's when you sort of understand what actually happens. And that's where everything starts to break. And you can do as much testing as you want when you, you know, when you actually go and out and like have humans using it, they do a lot of interesting things. And like I said, that heterogeneous environment this is deployed in leads all this instability. It's like really interesting to say, oh, what are these crazy problems that happen? And I'll give you an example of a bug that I think took us months to do. It was like, I think a carrier in the US rolled out IPv6 and didn't tell its customers it was. It did it in three cities. And, you know, we handled it fine for the most…
All right, Pete, that's all from me. Thank you so much for being today with us. And yeah, all the best with Peter.
Yeah, thank you so much. Very much appreciate being on here.